Data processing method and system for home industry drainage platform

Abstract

The invention relates to the technical field of data processing, and discloses a data processing method and system for a home industry drainage platform, and the method comprises the steps: obtaining and processing user behavior data, and obtaining a user preference cluster; extracting association vectors of product styles and material features based on the user preference cluster, and calculating similarity to determine a preliminary matching product set; high-frequency style and material combinations are extracted, and a recommended product list is obtained through screening; performing weighted fusion on the user behavior data to determine a recommendation weight set and performing sorting to obtain a product recommendation sequence; if the update difference of the user preference cluster exceeds the threshold value, updating the weight and adjusting the sequence; and generating a high-appeal abstract text based on the optimized sequence, determining personalized drainage content and synchronizing platform display logic. According to the method, the recommendation accuracy, the dynamic adaptability and the drainage effect can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to a data processing method and system for a household industry lead platform. BACKGROUND

[0002] The data processing of the household industry and the user lead platform play a key role in modern business. Through precise analysis of user behavior and product characteristics, it promotes personalized recommendations, significantly improves user experience and market competitiveness. With the deepening of digital transformation, the household industry's demand for data-driven marketing is increasingly urgent, and the platform needs to extract valuable information from massive data to achieve precise product recommendations and user lead.

[0003] In the prior art, many platforms rely on traditional statistical analysis or simple rule matching for product recommendations, such as recommending similar products based only on user browsing records, ignoring the hidden style preferences or material preferences in user preferences. These methods are difficult to cope with the multidimensionality of user behavior and the complexity of product attributes, and when dealing with high-dimensional, heterogeneous data, feature extraction is often not comprehensive, leading to recognition bias of style or material, and it is difficult to quickly adjust feature weights in dynamic data streams to adapt to changes in user preferences.

[0004] Therefore, the prior art cannot meet the needs of the household industry lead platform for precise and personalized recommendations, and there is a problem that the matching degree of the recommended results with the actual needs of the users is not high. SUMMARY

[0005] The present application provides a data processing method and system for a household industry lead platform to solve the problem that the matching degree of the recommended results with the actual needs of the users is not high.

[0006] In a first aspect, to solve the above technical problems, the present application provides a data processing method for a home industry drainage platform, comprising: obtaining user behavior data and processing to obtain a user preference cluster; based on the user preference cluster, extracting an associated vector of product style and material characteristics, calculating vector similarity, and determining a preliminary matched product set; extracting the style and material combination with the highest frequency of occurrence in the preliminary matched product set, filtering products from a pre-established product label database and evaluation feedback data set to obtain a recommended product list; weighting and fusing the user behavior data to determine a recommendation weight set and sort the recommended product list to obtain a product recommendation sequence; if the user preference cluster update difference exceeds a preset update threshold, updating the recommendation weight, adjusting the product recommendation sequence, and obtaining an optimized product recommendation sequence; based on the optimized product recommendation sequence, analyzing the style and material details of the goods, generating a high-attraction summary text, matching the user preference, and determining the final personalized drainage content; according to the final personalized drainage content, synchronizing the platform display logic in real time to obtain an updated configuration of the user interface.

[0007] In an optional implementation, the obtaining user behavior data and processing to obtain a user preference cluster comprises: obtaining user real-time browsing records, collection records, shopping cart product information, collecting interaction frequency and product category data, filtering the latest behavior data to obtain an original behavior data set; for the original behavior data set, data cleaning is performed to delete missing values, abnormal values and duplicate records to obtain the user behavior data; the user behavior data is grouped by using a K-means clustering algorithm, and the distance between user feature vectors is calculated to determine the user preference cluster.

[0008] In an optional implementation, the based on the user preference cluster, extracting an associated vector of product style and material characteristics, calculating vector similarity, and determining a preliminary matched product set comprises: obtaining product style and material characteristic data from the user preference cluster, performing data cleaning to remove noise data to obtain a cleaned preference data set; according to the cleaned preference data set, generating an embedding vector of product style and material characteristics by feature extraction to obtain a vector mapping set; if the embedding vector dimension in the vector mapping set meets a preset dimension threshold, comparing the vectors in the vector mapping set by calculating cosine similarity to obtain a similarity score matrix; according to the similarity score matrix, obtaining the products corresponding to the vectors with a score higher than a preset similarity threshold to determine a preliminary matched product set.

[0009] In an optional implementation, the extracting the most frequently occurring style and material combination in the preliminary matched product set, screening products from the pre-established product label database and evaluation feedback data set to obtain a recommended product list, includes: performing frequency statistics on the product features in the preliminary matched product set, extracting the high-frequency occurring style preference and material characteristic combination to obtain a refined feature set; if the matching degree of the style preference and the material characteristic in the refined feature set exceeds a preset matching degree threshold, screening products that meet the style preference and the material characteristic from the pre-established product label database and evaluation feedback data set to obtain a candidate product set; and determining a recommended product list by adjusting the recommendation priority using weighted sorting based on the candidate product set in combination with the browsing time data and click frequency records in the user behavior data.

[0010] In an optional implementation, the weighted fusion of the user behavior data, determination of a recommendation weight set, and sorting of the recommended product list to obtain a product recommendation sequence, includes: extracting user recent browsing time, click frequency, collection records, and shopping cart item information from the user behavior data; calculating the user behavior data using weighted fusion according to preset weight coefficients of each behavior indicator to obtain preference scores of different products, which constitute a recommendation weight set; and sorting the recommended product list according to the recommendation weight set to obtain the product recommendation sequence.

[0011] In an optional implementation, if the user preference cluster update difference exceeds a preset update threshold, updating the recommendation weight, adjusting the product recommendation sequence, and obtaining an optimized product recommendation sequence, includes: monitoring the update data of the user preference cluster in real time, calculating the feature difference value between the updated user preference cluster and the historical user preference cluster; if the feature difference value exceeds a preset update threshold, recalculating the preference score to generate a new recommendation weight; and re-sorting the product recommendation sequence based on the new recommendation weight to form an optimized product recommendation sequence.

[0012] In an optional implementation, based on the optimized product recommendation sequence, analyzing the style and material details of the goods, generating a high-attraction summary text, matching the user preference, and determining the final personalized lead content, includes: analyzing the product description text for the goods in the optimized product recommendation sequence, extracting the style and material details of the goods to obtain trend keywords and touch points; generating a high-attraction summary text based on the trend keywords and the touch points in combination with a preset scene association template; if the matching degree of the high-attraction summary text and the user preference cluster is lower than a preset matching threshold, adjusting the scene association content and the trend vocabulary to regenerate the summary text until the matching degree meets the standard to determine the final personalized lead content.

[0013] In a second aspect, the present application provides a data processing system for a home industry lead platform, comprising: a data acquisition module that acquires user behavior data and processes it to obtain a user preference cluster; a preliminary matching module that extracts an associated vector of product style and material characteristics based on the user preference cluster, calculates vector similarity, and determines a preliminary matched product set; a product screening module that extracts the style and material combination with the highest frequency of occurrence in the preliminary matched product set, screens products from a pre-established product label database and evaluation feedback data set, and obtains a recommended product list; a sequence generation module that performs weighted fusion on the user behavior data, determines a recommendation weight set and sorts the recommended product list, and obtains a product recommendation sequence; a sequence optimization module that updates the recommendation weight and adjusts the product recommendation sequence to obtain an optimized product recommendation sequence if the user preference cluster update difference exceeds a preset update threshold; a content output module that analyzes the style and material details of a product based on the optimized product recommendation sequence, generates a high-attraction summary text, matches user preferences, and determines the final personalized lead content; and an interface configuration module that synchronizes platform display logic in real time based on the final personalized lead content and obtains an updated configuration of the user interface.

[0014] In a third aspect, the present application also provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the data processing method for a home industry lead platform as described in any one of the above.

[0015] In a fourth aspect, the present application also provides a computer readable storage medium comprising a stored computer program, wherein the computer program, when executed, controls a device in which the computer readable storage medium is located to perform the data processing method for a home industry lead platform as described in any one of the above.

[0016] Compared with the prior art, the present application has the following beneficial effects: (1) The present application can accurately capture the potential preferences of users for home product style and material by acquiring user behavior data and a user preference cluster, extracting an associated vector of product style and material characteristics based on the user preference cluster, calculating similarity, and determining a preliminary matched product set, thereby solving the problem of disconnection between recommendations and user needs caused by the simple rule matching in the prior art and improving the accuracy of product recommendations; (2) The application determines a recommendation weight set by weighted fusion of user behavior data, sorts a product recommendation sequence from a recommended product list, and updates the recommendation weight and adjusts the sequence when the user preference cluster update difference exceeds a threshold, thereby achieving real-time response to dynamic user demand, overcoming the defect that it is difficult to quickly adjust feature weights to adapt to changes in user preferences in the prior art, and enhancing the dynamic adaptability of recommendations; (3) The application generates a high-attraction summary text based on the optimized product recommendation sequence, determines the final personalized lead content, and synchronizes the platform display logic, which can present the accurately recommended products in a way that fits the user's preferences, thereby improving the user's attention and acceptance of the lead content and effectively improving the user experience and conversion rate of the home industry lead platform. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a data processing method flow diagram for a home industry lead platform provided by the first embodiment of the application; Figure 2 is a data processing system structure diagram for a home industry lead platform provided by the second embodiment of the application. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the application.

[0019] Referring to Figure 1 The first embodiment of the application provides a data processing method for a home industry lead platform, including the following steps: S11, obtaining and processing user behavior data to obtain a user preference cluster; S12, based on the user preference cluster, extracting an associated vector of product style and material features, calculating vector similarity, and determining a preliminary matched product set; S13, extracting the style and material combination with the highest frequency of occurrence in the preliminary matched product set, and screening products from a pre-established product label database and evaluation feedback data set to obtain a recommended product list; S14, weighted fusion of the user behavior data to determine a recommendation weight set and sort the recommended product list to obtain a product recommendation sequence; S15, if the user preference cluster update difference exceeds a preset update threshold, updating the recommendation weight, adjusting the product recommendation sequence, and obtaining an optimized product recommendation sequence; S16, based on the optimized product recommendation sequence, parse the style and material details of the goods, generate high-attraction summary text, match user preferences, and determine the final personalized lead content; S17, according to the final personalized lead content, real-time synchronization is performed on the platform display logic to obtain the updated configuration of the user interface.

[0020] In step S11, user behavior data needs to be obtained and processed to obtain user preference clusters, including: obtaining user real-time browsing records, collection records, shopping cart commodity information, collecting interaction frequency and commodity category data, screening the latest behavior data to obtain the original behavior data set; for the original behavior data set, data cleaning is performed to delete missing values, outliers and duplicate records to obtain the user behavior data; the user behavior data is grouped by using the K-means clustering algorithm, the distance between user feature vectors is calculated, and the user preference cluster is determined.

[0021] It should be noted that the user behavior data refers to various interactive behavior records generated by the user on the platform, which can reflect the user's interest and demand tendency. Among them, the real-time browsing record includes the product page URL browsed by the user, the browsing timestamp and other information; the shopping cart commodity information covers the commodity data added to the shopping cart by the user, such as the name, model, price, and purchase time of the added shopping cart commodity; the interaction frequency refers to the number and frequency of user operations such as clicking, collecting, and adding to the shopping cart; the commodity category data is information classified according to certain classification standards, such as sofa, bed, dining table, etc. The user feature vector is a vector formed after quantizing these user behavior data, and each dimension represents the quantized value of a behavior feature. By calculating the distance between different user feature vectors, the similarity of user preferences can be determined.

[0022] In this step, first, the user's real-time browsing record and purchase history are obtained from the database of the home industry lead platform, and the user's interaction frequency and commodity category data are collected. In order to ensure the timeliness of the data, the behavior data in the recent period (such as the last 7 days) is selected to form the original behavior data set. Then, the original behavior data set is cleaned, which is a key link to ensure data quality. Specifically, check if there are missing values in the data, such as the user's browsing time being empty in a record; whether there are outliers, such as negative purchase amount or browsing time exceeding a reasonable range (such as more than 24 hours); and whether there are duplicate records, i.e. the same behavior of the same user to the same commodity is recorded multiple times. For these data that do not meet the requirements, they are deleted to obtain clean and accurate user behavior data.

[0023] In an implementation, the user behavior data is grouped using a K-means clustering algorithm. The K-means clustering algorithm is a commonly used unsupervised learning algorithm that divides the samples in a dataset into K different clusters, such that the samples within a cluster have a high similarity, while the samples between clusters have a low similarity. In this step, after the user behavior data is converted into user feature vectors, the similarity between users is measured by calculating the Euclidean distance between vectors, and then the users are divided into different user preference clusters, each cluster representing a group of users with similar preferences.

[0024] It is worth noting that filtering the latest behavior data can ensure that the analyzed data reflects the user's recent preferences, avoiding misjudgment of the user's current needs due to the use of outdated data. The data cleaning process can remove noise data and improve the accuracy of subsequent data analysis and processing. The clustering effect of the K-means clustering algorithm is affected by the selection of K value. In practical applications, the appropriate K value needs to be determined through multiple experiments based on the user scale and data characteristics of the platform to ensure that the division of user preference clusters is reasonable and effective.

[0025] For example, assume that on a home platform, user A has browsed the page of a fabric sofa 3 times in the last 7 days, with each browsing duration being 2 minutes, 3 minutes, and 5 minutes, and on August 12, user A purchased a modern and simple style fabric sofa; user B has browsed the page of a solid wood dining table 2 times in the last 7 days, and has collected a dining table of a Nordic style, without adding it to the shopping cart. After collecting these data, the latest data is filtered to form the original behavior dataset. In the data cleaning process, it is found that one of user A's browsing records is missing the browsing timestamp and is deleted; all of user B's data is complete and reasonable and is retained. Then, the behavior data of user A and user B is converted into feature vectors, assuming that the feature vector of user A is [fabric sofa browsed 3 times, added to shopping cart 1 time, click frequency 0.5 times / hour], and the feature vector of user B is [solid wood dining table browsed 2 times, collected 1 time, click frequency 0.3 times / hour]. Using the K-means clustering algorithm, set K = 2, calculate the Euclidean distance, and divide user A and other users with similar sofa preferences into one cluster, and divide user B and other users with similar dining table preferences into another cluster, thereby obtaining two different user preference clusters.

[0026] In step S12, it is necessary to extract the associated vectors of product style and material characteristics based on the user preference cluster, calculate the vector similarity, and determine the preliminary matched product set, including: obtaining product style and material characteristic data from the user preference cluster, performing data cleaning to remove noise data, and obtaining a cleaned preference data set; according to the cleaned preference data set, generating product style and material characteristic embedding vectors through feature extraction to obtain a vector mapping set; if the embedding vector dimension in the vector mapping set meets the preset dimension threshold, comparing the vectors in the vector mapping set by calculating the cosine similarity to obtain a similarity score matrix; according to the similarity score matrix, obtaining the products corresponding to the vectors with scores higher than the preset similarity threshold to determine the preliminary matched product set.

[0027] It should be noted that the product style feature refers to the design style attribute of the home product, such as Nordic style, modern minimalist style, Chinese classical style, etc.; the material feature is the material attribute used by the product, such as solid wood, metal, glass, cloth, etc. These features are important indicators to distinguish different home products and are important factors for users to select products. Embedding vector is a representation method that converts high-dimensional sparse feature data into low-dimensional dense vectors, which can capture the semantic association between features, so that similar features are closer in vector space. Cosine similarity is an index to measure the direction similarity of two vectors, and its value range is [-1, 1], the closer the value is to 1, the more similar the direction of the two vectors is.

[0028] In this step, first, the style and material characteristic data of the product concerned by the user are extracted from the user preference cluster obtained in step S11. These data may come from product label information, description text, etc. Then, the data is cleaned to remove noise data, such as incorrect style labels (mistaking modern minimalist style for Nordic style), ambiguous material descriptions (such as “composite material” without specific components), etc., to obtain a cleaned preference data set.

[0029] In an implementation, the cleaned preference dataset needs to be processed by feature extraction (such as Word2Vec, GloVe, etc. natural language processing model) to convert the style and material features of the product into embedding vectors, each feature corresponding to an embedding vector, thereby forming a vector mapping set. Then, check whether the dimension of the embedding vector in the vector mapping set meets the preset dimension threshold (such as 3 dimensions), if it meets, calculate the cosine similarity between the vectors. By calculating the cosine similarity between all vectors, a similarity score matrix is constructed, each element in the matrix represents the similarity score of the corresponding two vectors. Finally, according to the similarity score matrix, the products corresponding to the vectors with a score higher than the preset similarity threshold (such as 0.85) are screened out, these products have a high similarity with the features in the user preference cluster, thereby forming a preliminary matched product set.

[0030] It is worth noting that cleaning the product style and material feature data can ensure the accuracy of subsequent feature extraction and similarity calculation, and avoid deviation of the matching result caused by incorrect data. The dimension of the embedding vector needs to be determined according to the complexity of the feature and the amount of data, and a suitable dimension can better capture the association between features. The selection of the preset similarity threshold will affect the size of the preliminary matched product set, if the threshold is too high, it may result in too few matched products; if the threshold is too low, it may introduce too many irrelevant products, which needs to be adjusted according to the actual business needs.

[0031] For example, the style and material feature data of the products in the "Nordic style solid wood furniture preference cluster" are extracted, which may include features such as "Nordic style", "solid wood", "oak", "white", etc. During data cleaning, it is found that the material label in one data is "unknown", which is deleted; "Nordic" is uniformly corrected to "Nordic style" to ensure data consistency. Then, these features are converted into 3-dimensional embedding vectors through natural language processing, assuming that the Nordic style is mapped to vector [0.8, 0.2, 0.1], and the solid wood material is [0.6, 0.4, 0.3], forming a vector mapping set. It is found that the dimensions of these embedding vectors are all 3, which meet the preset dimension threshold. If a vector has only 2 dimensions due to data missing, it is excluded to ensure consistency. The cosine similarity between these vectors is calculated to obtain a similarity score matrix, assuming that the similarity score between "Nordic style" and the feature vector of a product labeled as "Nordic style solid wood dining table" is 0.92, which is higher than the preset similarity threshold 0.85, then this dining table is included in the preliminary matched product set.

[0032] In step S13, the combination of style and material with the highest frequency of occurrence in the preliminary matched product set needs to be extracted, and products are screened from the pre-established product label database and evaluation feedback dataset to obtain a recommended product list, including: frequency statistics of product features in the preliminary matched product set, extraction of high-frequency style preference and material characteristic combination, and obtaining a refined feature set; if the matching degree of style preference and material characteristic in the refined feature set exceeds the preset matching degree threshold, products conforming to the style preference and material characteristic are screened from the pre-established product label database and evaluation feedback dataset to obtain a candidate product set; according to the candidate product set, combined with the browsing time data and click frequency record in the user behavior data, a weighted sorting is adopted to adjust the recommendation priority, and a recommended product list is determined.

[0033] It should be noted that the frequency statistics of product features refer to counting the number of times of various style and material combinations in the preliminary matched product set. Through statistics, the most popular style and material combination in the user preference cluster can be found. The matching degree of style preference and material characteristic is an index to measure the rationality and coordination between the two. For example, the Nordic style is usually more coordinated with materials such as solid wood and cloth, and its matching degree is higher. The matching degree of Chinese classical style and metal material is relatively low.

[0034] The pre-established product label database is a database storing various label information of products, including style, material, color, size, etc. The evaluation feedback dataset contains user's evaluation, score, and use feedback of products, which can reflect the quality and user satisfaction of products.

[0035] In this step, first, the style and material characteristics of each product in the preliminary matched product set are extracted, and then the frequency of occurrence of each style and material combination is counted. For example, it is counted that the combination of “Nordic style + solid wood” appears 20 times, and the combination of “modern minimalist + cloth” appears 15 times, etc. The top several combinations with the highest frequency of occurrence are extracted to form a refined feature set. Then, the matching degree of each style preference and material characteristic in the refined feature set is calculated, the total number of occurrences of all material characteristic combinations under each style preference is counted, and the proportion of the former to the latter is calculated as the matching degree. If the matching degree exceeds the preset matching degree threshold (such as 0.8), products with these style and material labels are screened from the product label database, and reference is made to the evaluation feedback dataset to select products with higher user scores (such as scores not lower than 4.5) and better evaluations to form a candidate product set.

[0036] In an implementation, the candidate product set is weighted and ranked in combination with the browsing time data and click frequency records in the user behavior data. For example, the weight of the browsing time is set to 0.6, and the weight of the click frequency is set to 0.4. The comprehensive score of each product is calculated, the products are ranked from high to low according to the comprehensive score, the recommendation priority is adjusted, the products with high ranking are recommended first, and thus the recommended product list is determined.

[0037] It is worth noting that extracting high-frequency style and material combinations can focus on the most mainstream demand in the user preference cluster and improve the relevance of the recommendation. Referring to the evaluation feedback data to filter products can ensure that the recommended products have good quality and user recognition, and improve the user experience. The weight setting in the weighted ranking needs to be adjusted according to the business objectives and user behavior characteristics of the platform.

[0038] For example, there are 50 products in the preliminary matched product set. After counting the frequency of the style and material combinations of these products, it is found that the combination of "Nordic style + oak" appears 17 times, which is the highest frequency combination. It is included in the refined feature set. The matching degree of the combination is calculated. For example, the total appearance frequency of all material combinations under "Nordic style" is 20 times, and the matching degree of "Nordic style + oak" is 17 / 20=0.85, which exceeds the preset matching degree threshold 0.8. All products labeled as "Nordic style" and "oak" material are filtered from the product tag database, totaling 80 products. From the evaluation feedback data set, products with a user rating of not less than 4.5 are filtered, resulting in 60 products, which form the candidate product set. In combination with the user behavior data, the browsing time of a certain Nordic style oak dining table is 100 minutes, and the click frequency is 5 times. The browsing time of another Nordic style oak dining chair is 80 minutes, and the click frequency is 8 times. The weight of the browsing time is set to 0.6, and the weight of the click frequency is set to 0.4. The total browsing time is 500 minutes, and the total click frequency is 20 times. The comprehensive scores of the two products are calculated. The comprehensive score of the dining table is (100 / 500) x 0.6 + (5 / 20) x 0.4 = 0.22, and the comprehensive score of the dining chair is (80 / 500) x 0.6 + (8 / 20) x 0.4 = 0.256. According to the comprehensive score, the comprehensive score of the dining chair is higher, and the recommended order of the dining chair is adjusted to be in front of the dining table, and thus the recommended product list is obtained.

[0039] In step S14, the user behavior data needs to be weighted and fused to determine the recommendation weight set and sort the recommended product list to obtain the product recommendation sequence, including: extracting the user's recent browsing time, click frequency, collection record and shopping cart product information from the user behavior data; calculating the user's preference score for different products according to the preset weight coefficients of each behavior indicator to obtain the recommendation weight set; and sorting the recommended product list according to the recommendation weight set to obtain the product recommendation sequence.

[0040] It should be noted that the user's recent browsing time refers to the total time the user spends browsing a product page in the recent period of time, and the longer the time, the higher the user's interest in the product; the click frequency is the ratio of the number of times the user clicks on a product to the total number of clicks, reflecting the user's attention to the product; the collection record indicates that the user marks the product as a collection, which is a direct reflection of the user's interest in the product; and the shopping cart product information is the user's purchase record of the product or similar products, which can reflect the user's actual purchase preference. The weight coefficients of each behavior indicator are preset according to the influence of different behaviors on user preference, for example, the purchase behavior most directly reflects the user's preference, and the weight coefficient can be set higher; and the weight coefficient of the browsing time is relatively low.

[0041] In this step, first, the user's recent (such as the last 15 days) browsing time, click frequency, collection record (whether collected) and shopping cart product information (whether there is a purchase history, i.e., whether added to the shopping cart) of each product in the recommended product list are extracted from the user behavior data. Then, according to the preset weight coefficients of each behavior indicator (such as purchase weight 0.4, collection record weight 0.2, click frequency weight 0.2, and browsing time weight 0.2), the weight of each behavior indicator of each product is calculated to obtain the user's preference score for the product. For example, a product has a purchase history (marked as 1), a collection record (marked as 1), a click frequency of 0.3, and a browsing time of 10 minutes (standardized to 0.2), and the product's preference score is 1x0.4+1x0.2+0.3x0.2+0.2x0.2=0.7. The preference scores of all products constitute the recommendation weight set. Finally, the recommended product list is sorted according to the preference scores of each product in the recommendation weight set from high to low to obtain the product recommendation sequence, and the product with a higher preference score is placed in the front of the sequence.

[0042] It should be noted that the user behavior data is weighted and fused to consider the influence of multiple user behaviors on preference, avoid the deviation caused by a single behavior indicator, and make the obtained preference score more accurately reflect the user's real needs. The setting of the weight coefficient needs to be adjusted according to the actual data and business experience of the platform to improve the accuracy of the recommendation.

[0043] For example, there are 10 products in the recommended product list, and the user's recent behavior indicators for these 10 products are extracted from the user behavior data. The behavior indicators of product C are: has added to cart history, has collection record, click frequency 0.5, and browsing time 20 minutes (standardized to 0.4); the behavior indicators of product D are: no added to cart history, has collection record, click frequency 0.6, and browsing time 25 minutes (standardized to 0.5). The preset weight coefficients are: purchase history 0.4, collection record 0.2, click frequency 0.2, and browsing time 0.2. The preference score of product C is 1 x 0.4 + 1 x 0.2 + 0.5 x 0.2 + 0.4 x 0.2 = 0.78; the preference score of product D is 0 x 0.4 + 1 x 0.2 + 0.6 x 0.2 + 0.5 x 0.2 = 0.42. According to the preference score, product C will be ranked higher than product D in the product recommendation sequence.

[0044] In step S15, if the user preference cluster update difference exceeds the preset update threshold, the recommendation weight is updated, the product recommendation sequence is adjusted, and an optimized product recommendation sequence is obtained, including: monitoring the update data of the user preference cluster in real time, calculating the feature difference value between the updated user preference cluster and the historical user preference cluster; if the feature difference value exceeds the preset update threshold, the preference score is recalculated, and a new recommendation weight is generated; based on the new recommendation weight, the product recommendation sequence is reordered to form an optimized product recommendation sequence.

[0045] It should be noted that the update data of the user preference cluster refers to the data that causes the original user preference cluster features to change as the user continuously generates new behaviors (such as new browsing, adding to cart, collecting, etc.) on the platform. The feature difference value is a quantitative indicator for measuring the difference between the updated user preference cluster and the historical user preference cluster in features, which can be obtained by calculating the Euclidean distance or cosine distance of the feature vectors. The preset update threshold is a critical value for judging whether the user preference has changed significantly. When the feature difference value exceeds the threshold, it indicates that the user preference has changed significantly, and the recommendation strategy needs to be adjusted.

[0046] In this step, first, the user's behavior data changes are tracked in real time through the platform's data monitoring mechanism, and when the user generates new interaction behaviors (such as browsing new product categories, adding products with different styles than previous preferences, etc.), the user preference cluster is updated. Then, the feature difference value between the updated user preference cluster and the historical user preference cluster before the update is calculated. Specifically, the feature vectors of the two clusters can be compared, and the distance between the vectors can be calculated to obtain the difference value. If the feature difference value exceeds the preset update threshold (such as 0.3), it means that the user's preferences have changed significantly, and the preference scores of the user for each product in the recommended product list need to be recalculated. When recalculating, the same behavior indicators and weight coefficients as in step S14 are used, but the updated user behavior data is used to generate new recommendation weights. Finally, the original product recommendation sequence is reordered according to the new recommendation weights, with products that better meet the user's current preferences placed in the front, forming an optimized product recommendation sequence.

[0047] It is worth noting that real-time monitoring of user preference cluster update data can capture dynamic changes in user preferences in a timely manner, ensuring that recommended content can keep up with changes in user needs. Setting an update threshold can prevent frequent adjustments to recommendations due to occasional user behavior. Only when the preferences change significantly will the update be performed, ensuring the stability and effectiveness of the recommendations. The process of recalculating preference scores and adjusting the recommendation sequence ensures that the recommended results always align with the user's latest preferences, improving the user experience.

[0048] For example, a user preference cluster originally prefers furniture with a "modern minimalist + fabric" style, with a historical feature vector of [0.6, 0.3, 0.1] (corresponding to the weights of modern minimalist style, fabric material, and other features, respectively). After a period of time, the user frequently browses furniture with a "Nordic style + solid wood" style, adding multiple related behavior data, and the updated user preference cluster feature vector becomes [0.2, 0.2, 0.6] (corresponding to the weights of modern minimalist style, fabric material, and Nordic style + solid wood material, respectively). The Euclidean distance between the two feature vectors is calculated to obtain a feature difference value of 0.7, which exceeds the preset update threshold of 0.3, so the recommendation weights need to be updated. Recalculating the user's preference scores for each product in the recommended product list, it is found that the "Nordic style solid wood dining table" originally ranked lower has a significantly improved preference score, while the preference score of the "modern minimalist fabric sofa" has decreased. According to the new recommendation weights, the product recommendation sequence is reordered, with the "Nordic style solid wood dining table" placed in the front, forming an optimized product recommendation sequence.

[0049] In step S16, it is necessary to parse the style and material details of the goods based on the optimized product recommendation sequence, generate high-attraction summary text, match user preferences, and determine the final personalized lead content, including: for the goods in the optimized product recommendation sequence, parse the product description text, extract the style and material details of the goods, and obtain the trend keywords and touch points; according to the trend keywords and the touch points, combine the preset scene association template to generate high-attraction summary text; if the matching degree of the high-attraction summary text and the user preference cluster is lower than the preset matching threshold, adjust the scene association content and the trend vocabulary, and regenerate the summary text until the matching degree meets the standard, and determine the final personalized lead content.

[0050] It should be noted that the product description text is a detailed introduction to the goods on the platform, including style description (such as "simple and smooth lines, showing the Nordic style"), material description (such as "imported oak, hard texture"), and functional characteristics. Style and temperament refers to the design style and overall atmosphere embodied by the goods, such as the fresh and natural Nordic style, and the simple and efficient modern minimalist style; material details are specific descriptions of the materials used in the goods, such as the type of wood and the texture of fabric. Trend keywords are currently popular words related to style and design in the home industry, such as "minimalist" and "wabi-sabi aesthetics"; touch points refer to the tactile sensation brought by the material of the goods, such as "smooth" and "soft and comfortable".

[0051] The preset scene association template is a pre-designed text template for combining goods with life scenes, such as "living room scene" and "bedroom scene", which can help users better imagine the use effect of the goods in real life. The matching degree is an index to measure the degree of fit between high-attraction summary text and user preference clusters, which is obtained by calculating the degree of coincidence of keywords and user preference characteristics in the text.

[0052] In this step, first, for each product in the optimized product recommendation sequence, extract its product description text and use natural language processing techniques (such as word segmentation, keyword extraction, etc.) to analyze the text. From the analysis results, extract words that reflect the style and temperament of the goods (such as "Nordic style" and "minimalist design") and material details (such as "imported walnut" and "skin-friendly cotton and linen fabric"), and then extract trend keywords (such as "naturalism" and "luxury quality") and touch points (such as "warm and delicate" and "breathable and comfortable").

[0053] In one implementation, the appropriate preset scene association template (such as "living room leisure scene" template for a sofa) is selected according to the extracted trend keywords and touch selling points, the keywords and selling points are integrated into the template, and a high-attraction summary text is generated. For example, combined with the trend keywords "natural wood style" of "northern European style solid wood sofa", the touch selling points "warm and moist wood texture" and the "living room leisure scene" template, the summary text is generated: "This northern European style solid wood sofa, with natural wood style, shows a fresh atmosphere, and the warm and moist wood texture brings a comfortable experience, creating a leisure space for your living room."

[0054] In addition, the matching degree of the summary text with the user preference cluster needs to be calculated, and the user preference cluster contains the user's preference information for style, material, scene, etc. If the matching degree is lower than the preset matching threshold (such as 0.6), the scene association content (such as changing "living room leisure scene" to "family gathering scene") and trend vocabulary (such as changing "natural wood style" to "northern European minimalist style") are adjusted, the summary text is regenerated, and the matching degree is calculated again until the matching degree reaches the preset threshold. At this time, the summary text is the final personalized lead content.

[0055] It is worth noting that analyzing the product description text and extracting trend keywords and touch selling points can accurately grasp the core features of the product, laying a foundation for generating attractive lead content. Combined with the preset scene association template, the summary text can be more visually appealing, making it easier for users to feel involved and increasing their interest in the product. By adjusting the summary text multiple times to reach the matching degree threshold, the lead content can be highly consistent with the user's preferences, enhancing the lead effect.

[0056] For example, one of the optimized product recommendation sequences is a "Japanese wabi-sabi style ceramic vase", and its product description text is "made of coarse pottery, handmade, simple lines, showing Japanese wabi-sabi aesthetics, with a natural rough texture on the surface, suitable for displaying dried flowers, and adding a peaceful atmosphere to the home." Analyzing the text, extracting the style and tone as "Japanese wabi-sabi style" and "simple lines", and the material details as "coarse pottery", "handmade", and "rough texture", the trend keywords "wabi-sabi aesthetics" and "handmade art" and the touch selling point "natural roughness" are obtained. Combined with the "bedroom decoration scene" template, the summary text is generated: "This Japanese wabi-sabi style ceramic vase is made of handmade coarse pottery, with a natural rough texture that shows wabi-sabi aesthetics, adding a peaceful atmosphere to your bedroom, and is an excellent choice for displaying dried flowers."

[0057] The matching degree of the text with the user preference cluster is calculated, which shows that the user likes the style of "natural and simple", and the matching degree is 0.5, which is lower than the preset matching threshold 0.6. The scene association content is adjusted to "study room decoration scene", the trend vocabulary is changed to "natural and simple style", and the abstract text is regenerated: "This Japanese wabi-sabi style ceramic vase is made of hand-made coarse pottery, which brings a natural and rough touch. The natural and simple style fully embodies the wabi-sabi aesthetics, adding a quiet atmosphere to your study room. It looks more elegant when placed with dried flowers." The matching degree is calculated again as 0.7, reaching the preset threshold, and the text is determined as the final personalized lead content.

[0058] In step S17, it is necessary to synchronize the platform display logic in real time according to the final personalized lead content, and obtain the updated configuration of the user interface.

[0059] It should be noted that the personalized content preference in the user historical interaction data includes user behavior records such as click, dwell time, and sharing of lead content, which can reflect the user's acceptance and preference for different types of lead content. Platform display logic refers to the rules and methods of platform page display, including page layout (such as the arrangement of product cards), content display order (such as the position of lead content), visual elements (such as background color and font size), etc. Real-time synchronization means that the adjusted display logic is immediately applied to the platform to ensure that the user can see the updated interface in a timely manner.

[0060] In this step, first, analyze the user historical interaction data to understand the characteristics of the user's preference for different personalized content, such as the user's preference for clicking on lead content located at the top of the page, and the user's preference for a simple page layout. Combined with the characteristics of the final personalized lead content (such as content length, style, etc.), determine the adjustment direction of the platform display logic, such as placing the lead content in a more eye-catching position on the page, and using visual elements that match the content style. Then, according to the adjustment direction, set the display logic of the platform: in the page layout, expand the display area of the personalized lead content; in the content display order, place the lead content at the top of the page; in the visual elements, select colors (such as light color for Nordic style goods) and fonts (such as sans-serif font for minimalist style goods) that match the style of the goods.

[0061] In one implementation, the set display logic also needs to be synchronized in real time to the front-end system of the home industry lead platform through technical means to generate an updated configuration file of the user interface. When the user refreshes the page or logs in to the platform again, the system will load the updated configuration file, and the user interface will be updated to display the new interface containing the final personalized lead content.

[0062] It is worth noting that the adjustment direction of the display logic in combination with the user's personalized content preferences can make the platform interface update more in line with the user's usage habits and visual preferences, and improve the user's attention to the lead content. Adjusting the page layout, content display order, and visual elements of the display logic can highlight personalized lead content and enhance its appeal and recognition. Real-time synchronization of display logic and generation of user interface update configuration can ensure that users see the latest lead content in a timely manner, improving lead effectiveness and user experience.

[0063] For example, the final personalized lead content is a high-attraction summary text about "Nordic style solid wood desk". The user's historical interaction data shows that the user likes to click on the top of the page and the lead content with pictures, and prefers a light blue background. Therefore, the adjustment direction of the platform display logic is to place the lead content at the top of the page, with a high-definition picture of the desk, and a light blue background. According to the adjustment direction, the platform display logic is set as follows: in terms of page layout, the lead content area is set to the banner position at the top of the page, and the width occupies the entire page; in terms of content display order, the lead content is placed before all product recommendations; in terms of visual elements, the background color is set to light blue, and the font uses sans-serif. Real-time synchronization of these settings to the platform generates an updated configuration of the user interface. After the user refreshes the page, the light blue banner area at the top of the page displays the lead content and picture of the "Nordic style solid wood desk", which meets the user's usage habits and preferences.

[0064] In summary, the present application discloses a kind of data processing method for household industry drainage platform, including obtaining user behavior data and user preference cluster;Based on the user preference cluster, extract the associated vector of product style and material characteristics, calculate vector similarity, determine the preliminary matched product set;Extract the highest frequency style and material combination in the preliminary matched product set, filter products from pre-established product label database and evaluation feedback data set, obtain recommended product list;Weighted fusion is carried out to the user behavior data, determines the recommendation weight set and sorts the recommended product list, obtains product recommendation sequence;If the user preference cluster update difference exceeds the preset update threshold, update recommendation weight, adjust the product recommendation sequence, obtain optimized product recommendation sequence;Based on the optimized product recommendation sequence, analyze the style of commodity and material details, generate high attractive abstract text, match user preference, determine the final personalized drainage content;According to the final personalized drainage content, real-time synchronization is carried out to platform display logic, obtains the update configuration of user interface.The present application can accurately capture the potential preference of user to household product style and material by obtaining user behavior data and user preference cluster, based on user preference cluster extracting the associated vector of product style and material characteristics and calculating similarity, determining the preliminary matched product set, solve the problem that only simple rule matching is relied on in prior art, leading to the disconnection between recommendation and user demand, improve the accuracy of product recommendation.And by weighted fusion to user behavior data to determine the recommendation weight set, the recommended product list is sorted to obtain the product recommendation sequence, and when the user preference cluster update difference exceeds threshold, the recommendation weight is updated and the sequence is adjusted, the real-time response to user dynamic demand is realized, the defect that it is difficult to quickly adjust feature weight to adapt to user preference change in prior art is overcome, and the dynamic adaptability of recommendation is enhanced.In addition, the present application generates high attractive abstract text based on optimized product recommendation sequence, determines the final personalized drainage content and synchronizes platform display logic, can present the accurately recommended product in a way that fits user preference, improves the attention and acceptance of user to drainage content, effectively improves the user experience and conversion rate of household industry drainage platform.

[0065] Reference Figure 2The second embodiment of the present application provides a data processing system for a household industry lead platform, comprising: a data acquisition module, which acquires user behavior data and processes the user behavior data to obtain a user preference cluster; a preliminary matching module, which extracts an associated vector of product style and material characteristics based on the user preference cluster, calculates a vector similarity, and determines a preliminary matched product set; a product screening module, which extracts the style and material combination with the highest occurrence frequency in the preliminary matched product set, screens products from a pre-established product label database and an evaluation feedback data set, and obtains a recommended product list; a sequence generation module, which performs weighted fusion on the user behavior data, determines a recommendation weight set, and sorts the recommended product list to obtain a product recommendation sequence; a sequence optimization module, which updates the recommendation weight and adjusts the product recommendation sequence to obtain an optimized product recommendation sequence if the user preference cluster update difference exceeds a preset update threshold; a content output module, which analyzes the style and material details of a product based on the optimized product recommendation sequence, generates a high-attraction summary text, matches user preferences, and determines final personalized lead content; and an interface configuration module, which performs real-time synchronization on platform display logic based on the final personalized lead content to obtain an updated configuration of a user interface.

[0066] It should be noted that the data processing system for a household industry lead platform provided by the embodiments of the present application is used to execute all process steps of the data processing method for a household industry lead platform of the above-mentioned embodiments, and the working principles and beneficial effects of the two are one-to-one corresponding, thus no longer being described.

[0067] The embodiments of the present application also provide an electronic device. The electronic device comprises a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a data processing program for a household industry lead platform. The processor implements the steps in each of the above-mentioned data processing methods for a household industry lead platform when executing the computer program, such as the step S11 shown in the above-mentioned data processing method for a household industry lead platform. Figure 1 Alternatively, the processor implements the functions of each module / unit in each of the above-mentioned device embodiments when executing the computer program, such as the sequence generation module.

[0068] For example, the computer program can be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present application. The one or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program in the electronic device.

[0069] The electronic device can be a computing device such as a desktop computer, a notebook computer, a palm computer, a smart tablet, etc. The electronic device can include, but is not limited to, a processor, a memory. Those skilled in the art can understand that the above components are only examples of the electronic device and do not constitute a limitation on the electronic device, and the electronic device can include more or fewer components than the above, or combine certain components, or different components, for example, the electronic device can also include an input / output device, a network access device, a bus, etc.

[0070] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The processor is the control center of the electronic device, and connects various parts of the electronic device through various interfaces and lines.

[0071] The memory can be used to store the computer program and / or modules, and the processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required by a function (such as a sound playing function, an image playing function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory device.

[0072] The modules / units integrated in the electronic device, if realized in the form of software function units and sold or used as independent products, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. The computer program can implement the steps of each method embodiment when executed by a processor. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer readable medium can include any entity or device, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. that can carry the computer program code. It should be noted that the contents included in the computer readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction, for example, in some jurisdictions, according to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

[0073] It should be noted that the above-described device embodiments are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. In addition, the connection relationship between the modules in the device embodiment provided by the present application indicates that there is a communication connection between them, which can be realized as one or more communication buses or signal lines. Those skilled in the art can understand and implement it without creative labor.

[0074] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are only for the specific embodiments of the present application and do not limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A data processing method for a home furnishing industry traffic diversion platform, characterized in that: include: Obtain user behavior data and process it to obtain user preference clusters; based on the user preference clusters, extract the association vectors of product style and material features, calculate vector similarity, and determine a preliminary matching product set; extract the style and material combinations that appear most frequently in the preliminary matching product set, filter products from a pre-established product tag database and evaluation feedback data set, and obtain a recommended product list; perform weighted fusion on the user behavior data, determine a recommendation weight set, and sort the recommended product list to obtain a product recommendation sequence; if the update difference of the user preference cluster exceeds a preset update threshold, update the recommendation weight, adjust the product recommendation sequence, and obtain an optimized product recommendation sequence; based on the optimized product recommendation sequence, analyze the style and material details of the product, generate a highly attractive summary text, match user preferences, and determine the final personalized drainage content; based on the final personalized drainage content, synchronize the platform display logic in real time to obtain an updated configuration of the user interface.

2. The data processing method for the home furnishing industry traffic diversion platform according to claim 1, characterized in that: The method of obtaining and processing user behavior data to obtain user preference clusters includes: obtaining users' real-time browsing records, favorite records, and shopping cart item information, collecting interaction frequency and item category data, filtering the latest behavior data, and obtaining an original behavior data set; performing data cleaning on the original behavior data set to delete missing values, outliers, and duplicate records to obtain the user behavior data; and grouping the user behavior data using a K-means clustering algorithm, and determining user preference clusters by calculating the distance between user feature vectors.

3. The data processing method for the home furnishing industry traffic diversion platform according to claim 1, characterized in that: The method of extracting association vectors of product styles and material features based on the user preference cluster, calculating vector similarity, and determining a preliminarily matched product set includes: obtaining product style and material feature data from the user preference cluster, performing data cleaning to remove noise data, and obtaining a cleaned preference data set; generating embedding vectors of product styles and material features through feature extraction based on the cleaned preference data set, and obtaining a vector mapping set; if the dimension of the embedded vectors in the vector mapping set meets a preset dimension threshold, comparing the vectors in the vector mapping set by calculating cosine similarity to obtain a similarity score matrix; and obtaining products corresponding to vectors with scores higher than the preset similarity threshold based on the similarity score matrix, and determining a preliminarily matched product set.

4. The data processing method for the home furnishing industry traffic diversion platform according to claim 1, characterized in that: The extracting of the style and material combinations that appear most frequently in the preliminary matching product set, screening products from a pre-established product label database and evaluation feedback data set, and obtaining a recommended product list includes: performing frequency statistics on product features in the preliminary matching product set, extracting frequently appearing style preferences and material characteristic combinations, and obtaining a refined feature set; if the matching degree between the style preferences and material characteristics in the refined feature set exceeds a preset matching degree threshold, screening products that meet the style preferences and material characteristics from a pre-established product label database and evaluation feedback data set to obtain a candidate product set; based on the candidate product set, combined with the browsing time data and click frequency records in the user behavior data, using weighted sorting to adjust the recommendation priority and determine a recommended product list.

5. The data processing method for the home furnishing industry traffic diversion platform according to claim 1, characterized in that: The weighted fusion of the user behavior data, determination of a recommendation weight set, and sorting of the recommended product list to obtain a product recommendation sequence includes: extracting the user's recent browsing time, click frequency, favorite records, and shopping cart item information from the user behavior data; calculating the user behavior data using weighted fusion according to preset weight coefficients of each behavior indicator to obtain the user's preference scores for different products to form a recommendation weight set; and sorting the recommended product list according to the recommendation weight set to obtain the product recommendation sequence.

6. The data processing method for the home furnishing industry traffic diversion platform according to claim 1, characterized in that: If the update difference of the user preference cluster exceeds a preset update threshold, the recommendation weight is updated, the product recommendation sequence is adjusted, and an optimized product recommendation sequence is obtained, including: real-time monitoring of the update data of the user preference cluster, and calculating the feature difference value between the updated user preference cluster and the historical user preference cluster; if the feature difference value exceeds the preset update threshold, the preference score is recalculated to generate a new recommendation weight; and the product recommendation sequence is reordered based on the new recommendation weight to form an optimized product recommendation sequence.

7. The data processing method for the home furnishing industry traffic diversion platform according to claim 1, characterized in that: The method of analyzing the style and material details of the products based on the optimized product recommendation sequence, generating a highly attractive summary text, matching user preferences, and determining the final personalized traffic-generating content includes: analyzing the product description text for the products in the optimized product recommendation sequence, extracting the style and material details of the products, and obtaining trendy keywords and tactile selling points; generating a highly attractive summary text based on the trendy keywords and the tactile selling points, combined with a preset scene association template; if the matching degree of the highly attractive summary text and the user preference cluster is lower than a preset matching threshold, adjusting the scene association content and trendy vocabulary, and regenerating the summary text until the matching degree meets the standard, and determining it as the final personalized traffic-generating content.

8. A data processing system for a home furnishing industry traffic diversion platform, characterized in that: include: The data acquisition module acquires and processes user behavior data to obtain user preference clusters. The preliminary matching module extracts the correlation vectors of product style and material characteristics based on the user preference clusters, calculates the vector similarity, and determines the preliminary matching product set. A product screening module extracts the most frequently appearing style and material combinations from the pre-matched product set, filters products from a pre-established product tag database and evaluation feedback dataset, and obtains a recommended product list. A sequence generation module performs weighted fusion on the user behavior data, determines a recommendation weight set, and sorts the recommended product list to obtain a product recommendation sequence. A sequence optimization module updates the recommendation weights and adjusts the product recommendation sequence to obtain an optimized product recommendation sequence if the update difference of the user preference cluster exceeds a preset update threshold; The content output module analyzes the style and material details of the products based on the optimized product recommendation sequence, generates highly attractive summary text, matches user preferences, and determines the final personalized traffic-generating content; The interface configuration module synchronizes the platform display logic in real time according to the final personalized traffic diversion content to obtain an updated configuration of the user interface.

9. An electronic device, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the data processing method for the home industry traffic diversion platform as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein, when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the data processing method for the home furnishing industry drainage platform as described in any one of claims 1 to 7.

Images